Zobrazeno 1 - 10
of 579
pro vyhledávání: '"Burlachenko A"'
Federated Learning (FL) is an emerging paradigm that enables intelligent agents to collaboratively train Machine Learning (ML) models in a distributed manner, eliminating the need for sharing their local data. The recent work (arXiv:2106.02969) intro
Externí odkaz:
http://arxiv.org/abs/2410.08760
Autor:
Salishcheva Olesya, Burlachenko Anastasia, Tarasova Yuliya, Moldagulova Natalia, Yustratov Vladimir
Publikováno v:
BIO Web of Conferences, Vol 64, p 01003 (2023)
Biodegradation is a sustainable and efficient method for removing organic pollutants from the aquatic environment. We studied the biological purification of aqueous solutions from betaine organic matter under the action of bacterial strains of the ge
Externí odkaz:
https://doaj.org/article/a354c7537d744bc5bcc17e44c0d56e98
Publikováno v:
E3S Web of Conferences, Vol 401, p 01040 (2023)
In the article are presented the results of the state analysis of low-pointed tubular culverts widely used in land reclamation and fisheries construction, the lower pool of which is made in the form of an expanding socket with a central angle of 300
Externí odkaz:
https://doaj.org/article/015574c41d544a7bbd9ae1333710b111
Publikováno v:
E3S Web of Conferences, Vol 365, p 03048 (2023)
Issues related to the design and hydraulic calculations of the transit part of a closed tubular water supply structure of the siphon type are considered. The operating conditions of a siphon with rectangular pipes in the transition mode are estimated
Externí odkaz:
https://doaj.org/article/73c6bfd818ff495bb85e6e57746d09bc
Autor:
Malinovskii, Vladimir, Mazur, Denis, Ilin, Ivan, Kuznedelev, Denis, Burlachenko, Konstantin, Yi, Kai, Alistarh, Dan, Richtarik, Peter
There has been significant interest in "extreme" compression of large language models (LLMs), i.e., to 1-2 bits per parameter, which allows such models to be executed efficiently on resource-constrained devices. Existing work focused on improved one-
Externí odkaz:
http://arxiv.org/abs/2405.14852
Error Feedback (EF) is a highly popular and immensely effective mechanism for fixing convergence issues which arise in distributed training methods (such as distributed GD or SGD) when these are enhanced with greedy communication compression techniqu
Externí odkaz:
http://arxiv.org/abs/2402.10774
Publikováno v:
Proceedings of the 4th International Workshop on Distributed Machine Learning December 2023
Traditional AI methodologies necessitate centralized data collection, which becomes impractical when facing problems with network communication, data privacy, or storage capacity. Federated Learning (FL) offers a paradigm that empowers distributed AI
Externí odkaz:
http://arxiv.org/abs/2312.02074
We provide the first proof that gradient descent $\left({\color{green}\sf GD}\right)$ with greedy sparsification $\left({\color{green}\sf TopK}\right)$ and error feedback $\left({\color{green}\sf EF}\right)$ can obtain better communication complexity
Externí odkaz:
http://arxiv.org/abs/2305.15264
Publikováno v:
Литературный факт, Iss 2 (32), Pp 183-203 (2024)
For the first time, the article examines the critical reception of the poetry of Georgy Ivanovich Rybintsev (1885–1944), an “unknown poet” who unwittingly became involved in the literary struggle of 1913–1914. Rybintsev’s first collection o
Externí odkaz:
https://doaj.org/article/10489d45fc994817b4f7f66816ab959b
Publikováno v:
Vestnik MGSU, Vol 19, Iss 3, Pp 456-468 (2024)
Introduction. The existing approaches to management decision-making in construction production give the opportunity to select options only for a certain stage of the capital construction object life cycle (LC CCO), not considering their effectiveness
Externí odkaz:
https://doaj.org/article/1b05d0b7a3b748748a8a4b4f17d615de